PROJECT SUMMARY/ABSTRACT Dr. McCoy is an endocrinologist and primary care physician specializing in the care for patients with diabetes and other chronic health conditions. Her long-term goal is to become an independent researcher and leader in evaluating, improving, and individualizing diabetes care through effective use of big data completed by qualitative insights from patients and those involved in their care. The short-term training goals of this proposal are to acquire: 1) proficiency in statistics and data science; 2) experience in qualitative research; 3) skills in clinical trial design; and 4) communication and leadership skills to lead multi-disciplinary research. The proposal's research will be conducted at Mayo Clinic, which has a strong history of training physician scientists and a well-developed infrastructure for education and research. Dr. McCoy has access to an ideal set of data assets, the OptumLabs Data Warehouse and the Kaiser Permanente Northern California Diabetes Registry, and the support of an exceptional team of mentors and advisors who are national leaders in diabetes outcomes and health care delivery research, shared decision-making, qualitative research, and data science. Optimization of glycemic control, while avoiding severe hypoglycemia and hyperglycemia, is the cornerstone of diabetes management. Severe hypoglycemia and hyperglycemia are often preventable, yet continue to incur substantial morbidity, psychological distress, impaired quality of life, and economic burden. There are no validated tools to predict these events and as a result clinicians lack a practical and reliable means to identify high risk patients. The goal of Dr. McCoy's work is to address this critical gap in diabetes management. In Aim 1, she will use large database analysis and novel analytic methods to characterize the patterns of severe hypoglycemia and hyperglycemia among adults with diabetes in the U.S. In Aim 2, she will build on Aim 1 to develop and validate computationally efficient concurrent risk prediction models for severe hypoglycemia and hyperglycemia that could be used in clinical encounters. In Aim 3, she will directly engage patients and their clinicians in conversation about severe hypoglycemia /hyperglycemia risk in order to better understand how information about severe hypoglycemia/ hyperglycemia risk is perceived, interpreted, and used. These studies will serve as foundation for two R01 applications to be submitted during Years 4 and 5, and will advance the science and practice of personalized diabetes care through integration of data science and qualitative methods for the purpose of understanding, predicting, and ultimately preventing severe hypoglycemic/hyperglycemic events. Dr. McCoy has a proven track record of scientific productivity and innovation, and a strong foundation in using secondary data for health services and outcomes research in diabetes. In summary, the training, mentoring, and research proposed here are essential for Dr. McCoy's career development as she develops research independence and expertise in the crucial interface between patients and their clinical teams, evidence-based medicine, and big data.